Identification of whole-brain network modules based on a large scale Granger Causality approach

Area of Science:

• Neuroscience
• Computational Neuroscience
• Network Science

Background:

• Analyzing high-resolution neurophysiological data for human brain functional networks presents significant computational challenges.
• Existing methods struggle with the scale and complexity of spatially high-resolved brain data.

Purpose of the Study:

• To introduce a novel multivariate linear Granger Causality approach with dimension reduction for large-scale brain network computation.
• To integrate network theory module detection for identifying functionally associated brain areas within high-dimensional directed networks.

Main Methods:

• Developed a multivariate linear Granger Causality method with embedded dimension reduction.
• Applied module detection algorithms from network theory to analyze connectivity patterns.
• Validated the methodology using synthetic data with known module properties.
• Demonstrated applicability on resting-state functional MRI (fMRI) data.

Main Results:

• The proposed method enables the computation of brain networks at a large scale.
• Module detection effectively identifies functionally associated brain areas within the computed networks.
• Successful verification with synthetic data and demonstration on clinical resting-state fMRI data.

Conclusions:

• The integrated approach offers a powerful solution for analyzing complex, high-dimensional neurophysiological data.
• This methodology enhances the identification of functional brain networks and connectivity patterns.
• The approach is beneficial for both research and clinical applications involving brain imaging data.